Arabic speech recognition, classification and text-to-speech.

Last update: Dec 27, 2022

Overview

klaam

Arabic speech recognition, classification and text-to-speech using many advanced models like wave2vec and fastspeech2. This repository allows training and prediction using pretrained models.

Usage

from klaam import SpeechClassification
model = SpeechClassification()
model.classify(wav_file)

from klaam import SpeechRecognition
model = SpeechRecognition()
model.transcribe(wav_file)

from klaam import TextToSpeech
model = TextToSpeech()
model.synthesize(sample_text)

There are two avilable models for recognition trageting MSA and egyptian dialect . You can set any of them using the lang attribute

 from klaam import SpeechRecognition
 model = SpeechRecognition(lang = 'msa')
 model.transcribe('file.wav')

Datasets

Dataset	Description	link
MGB-3	Egyptian Arabic Speech recognition in the wild. Every sentence was annotated by four annotators. More than 15 hours have been collected from YouTube.	requires registeration here
ADI-5	More than 50 hours collected from Aljazeera TV. 4 regional dialectal: Egyptian (EGY), Levantine (LAV), Gulf (GLF), North African (NOR), and Modern Standard Arabic (MSA). This dataset is a part of the MGB-3 challenge.	requires registeration here
Common voice	Multlilingual dataset avilable on huggingface	here.
Arabic Speech Corpus	Arabic dataset with alignment and transcriptions	here.

Models

We currently support four models, three of them are avilable on transformers.

Language	Description	Source
Egyptian	Speech recognition	wav2vec2-large-xlsr-53-arabic-egyptian
Standard Arabic	Speech recognition	wav2vec2-large-xlsr-53-arabic
EGY, NOR, LAV, GLF, MSA	Speech classification	wav2vec2-large-xlsr-dialect-classification
Standard Arabic	Text-to-Speech	fastspeech2

Example Notebooks

Name	Description	Notebook
Demo	Classification, Recongition and Text-to-speech in a few lines of code.
Demo with mic	Audio Recongition and classification with recording.

Training

The scripts are a modification of jqueguiner/wav2vec2-sprint.

classification

This script is used for the classification task on the 5 classes.

python run_classifier.py \
   --model_name_or_path="facebook/wav2vec2-large-xlsr-53" \
   --output_dir=/path/to/output \
   --cache_dir=/path/to/cache/ \
   --freeze_feature_extractor \
   --num_train_epochs="50" \
   --per_device_train_batch_size="32" \
   --preprocessing_num_workers="1" \
   --learning_rate="3e-5" \
   --warmup_steps="20" \
   --evaluation_strategy="steps"\
   --save_steps="100" \
   --eval_steps="100" \
   --save_total_limit="1" \
   --logging_steps="100" \
   --do_eval \
   --do_train \

Recognition

This script is for training on the dataset for pretraining on the egyption dialects dataset.

python run_mgb3.py \
    --model_name_or_path="facebook/wav2vec2-large-xlsr-53" \
    --output_dir=/path/to/output \
    --cache_dir=/path/to/cache/ \
    --freeze_feature_extractor \
    --num_train_epochs="50" \
    --per_device_train_batch_size="32" \
    --preprocessing_num_workers="1" \
    --learning_rate="3e-5" \
    --warmup_steps="20" \
    --evaluation_strategy="steps"\
    --save_steps="100" \
    --eval_steps="100" \
    --save_total_limit="1" \
    --logging_steps="100" \
    --do_eval \
    --do_train \

This script can be used for Arabic common voice training

python run_common_voice.py \
    --model_name_or_path="facebook/wav2vec2-large-xlsr-53" \
    --dataset_config_name="ar" \
    --output_dir=/path/to/output/ \
    --cache_dir=/path/to/cache \
    --overwrite_output_dir \
    --num_train_epochs="1" \
    --per_device_train_batch_size="32" \
    --per_device_eval_batch_size="32" \
    --evaluation_strategy="steps" \
    --learning_rate="3e-4" \
    --warmup_steps="500" \
    --fp16 \
    --freeze_feature_extractor \
    --save_steps="10" \
    --eval_steps="10" \
    --save_total_limit="1" \
    --logging_steps="10" \
    --group_by_length \
    --feat_proj_dropout="0.0" \
    --layerdrop="0.1" \
    --gradient_checkpointing \
    --do_train --do_eval \
    --max_train_samples 100 --max_val_samples 100

Text To Speech

We use the pytorch implementation of fastspeech2 by ming024. The procedure is as follows

Download the dataset

wget http://en.arabicspeechcorpus.com/arabic-speech-corpus.zip 
unzip arabic-speech-corpus.zip

Create multiple directories for data

mkdir -p raw_data/Arabic/Arabic preprocessed_data/Arabic/TextGrid/Arabic
cp arabic-speech-corpus/textgrid/* preprocessed_data/Arabic/TextGrid/Arabic

Prepare metadata

import os 
base_dir = '/content/arabic-speech-corpus'
lines = []
for lab_file in os.listdir(f'{base_dir}/lab'):
  lines.append(lab_file[:-4]+'|'+open(f'{base_dir}/lab/{lab_file}', 'r').read())


open(f'{base_dir}/metadata.csv', 'w').write(('\n').join(lines))

Clone my fork

git clone --depth 1 https://github.com/zaidalyafeai/FastSpeech2
cd FastSpeech2
pip install -r requirements.txt

Prepare alignments and prepreocessed data

python3 prepare_align.py config/Arabic/preprocess.yaml
python3 preprocess.py config/Arabic/preprocess.yaml

Unzip vocoders

unzip hifigan/generator_LJSpeech.pth.tar.zip -d hifigan
unzip hifigan/generator_universal.pth.tar.zip -d hifigan

Start training

python3 train.py -p config/Arabic/preprocess.yaml -m config/Arabic/model.yaml -t config/Arabic/train.yaml

Arabic speech recognition, classification and text-to-speech.

Related tags

Overview

klaam

Usage

Datasets

Models

Example Notebooks

Training

classification

Recognition

Text To Speech

Owner

ARBML

Neural network models for joint POS tagging and dependency parsing (CoNLL 2017-2018)

Community and sentiment analysis based on tweets

SDL: Synthetic Document Layout dataset

Open-World Entity Segmentation

A BERT-based reverse-dictionary of Korean proverbs

Jarvis is a simple Chatbot with a GUI capable of chatting and retrieving information and daily news from the internet for it's user.

📜 GPT-2 Rhyming Limerick and Haiku models using data augmentation

A CRM department in a local bank works on classify their lost customers with their past datas. So they want predict with these method that average loss balance and passive duration for future.

The ibet-Prime security token management system for ibet network.

Implementation of some unbalanced loss like focal_loss, dice_loss, DSC Loss, GHM Loss et.al

Official source for spanish Language Models and resources made @ BSC-TEMU within the "Plan de las Tecnologías del Lenguaje" (Plan-TL).

BERT score for text generation

Research Code for NeurIPS 2020 Spotlight paper "Large-Scale Adversarial Training for Vision-and-Language Representation Learning": UNITER adversarial training part

2021语言与智能技术竞赛：机器阅读理解任务

基于Transformer的单模型、多尺度的VAE模型

A simple chatbot based on chatterbot that you can use for anything has basic features

A Japanese tokenizer based on recurrent neural networks

a chinese segment base on crf

Shared code for training sentence embeddings with Flax / JAX

Bidirectional Variational Inference for Non-Autoregressive Text-to-Speech (BVAE-TTS)